Clinical and environmental predictors of preterm labor.

OBJECTIVE: To determine the magnitude of risk for preterm labor associated with specific clinical and environmental factors.
METHODS: Using a case-control design, 266 women with preterm labor and 512 controls were interviewed and their medical records reviewed. Crude and adjusted odds ratios were calculated for each risk factor. Population-attributable risks were estimated.
RESULTS: Third-trimester bleeding, twin gestation, and chorioamnionitis at presentation were strongly associated with preterm labor (odds ratios 11.2-48.3). A history of a prior preterm delivery, vaginal bleeding in the first or second trimester, maternal diethylstilbestrol exposure, uterine anomalies, and urinary tract infection during pregnancy were associated to a lesser extent (odds ratios 1.6-5.4), as were cigarette smoking and drug use (odds ratios 2.0 and 3.0). Cases who had preterm labor preceded by premature rupture of the membranes had a substantially higher risk of preterm labor if chorioamnionitis, vaginal bleeding early in pregnancy, or urinary tract infection was present. By contrast, women who had intact membranes at the onset of preterm labor carried higher risk when twin gestation, placental abruption, or uterine anomaly was present. The highest population-attributable risks for preterm labor were found in patients with a twin gestation or third-trimester bleeding.
CONCLUSIONS: Programs to reduce the preterm delivery rate should consider the attributable risks for the factors they are intended to modify. The attributable risks we obtained suggest that medical strategies to reduce the impact of the clinical variables, especially multiple gestation, and educational programs to decrease smoking and drug use should reduce the preterm delivery rate.